pokemove

Inspiration

We wanted to make a project that involved physical exercise. But we didn't want it to be plain and boring so we added 2 enhancements: gamification and machine learning. Pokemove is a fun, innovative take on yoga.

What it does

Pokemove allows users to engage in yoga training sessions, guided by ML based annotations to help correct their form so they can copy the yoga pros. Each time they hit and keep the correct pose for 10 seconds, they're rewarded with a Pokemon. Gotta catch em all!

How we built it

React, Express, Tensorflow/Posenet, Node.js, No.sleep,

Challenges we ran into

Timezone/scheduling/availability, first time using posenet

Accomplishments that we're proud of

We think we built a very unique app that could have a high positive impact on users' physical health

What we learned

Architecture design, using machine learning specifically posenet, real time video detection

What's next for pokemove

We want to flesh out the game further with improved design and animations. We also want to offer levels so beginners and advanced yoga practitioners can choose their course. We eventually want to build out a network so people can trade Pokemon with their friends.

Github: https://github.com/shreyjay0/pokemove https://github.com/gracew99/pokemove-backend

Built With

• express.js
• node.js
• react
• tensorflow